Selected Variables

base: Code of the patient
covariates:
- Age
- Gender
- Prior Spine Surgery
- ASA classification
- Decompression
- Osteotomy
- 3CO
- SPOs
- BMI_First Visit
- Tobacco use_First Visit
- Osteoporosis / osteopenia
- Levels Previously operated - Lower
- LGap
- RLL
- Number of Interbody Fusions
- 'Posterior Instrumented Fusion: Upper / Lower Levels'
- Alif
- LL-Lordosis Difference
outcomes_ql:
- 2Y. ODI - Score (%)
- 2Y. SRS22 - SRS Subtotal score
- 2Y. SF36 - MCS
- 2Y. SF36 - PCS
outcomes_radiology:
- 6W. Major curve Cobb angle
- 1Y. Major curve Cobb angle
- 6W. T1 Sagittal Tilt
- 1Y. T1 Sagittal Tilt
- 6W. Sagittal Balance
- 1Y. Sagittal Balance
- 6W. Global Tilt
- 1Y. Global Tilt
- 6W. Lordosis (top of L1-S1)
- 1Y. Lordosis (top of L1-S1)
- 6W. LGap
- 1Y. LGap
- 6W. Pelvic Tilt
- 1Y. Pelvic Tilt
predictive:
- Weight (kgs)_First Visit
- Height (cm)_First Visit
- Total surgical time st1+st2+st3
- Osteotomy
- Alcohol/drug abuse
- Anemia or other blood disorders
- Osteoarthritis
- Mild vascular
- Depression / anxiety
- Diabetes with end organ damage
- Cardiac
- Hypertension
- Chronic pulmonary disease
- Nervous system disorders
- Renal
- Peripheral vascular disease
- Psychiatric / Behavioral
- Peptic ulcer
- Bladder incontinence
- Bowel incontinence
- Leg weakness
- Loss of balance
- NRS back - Leg pain - Average
- Tobacco use_First Visit
- Years with spine problems
- ODI - Score (%)_First Visit
- SRS22 - SRS Total score_First Visit
- SF36 - PCS_First Visit
- SF36 - MCS_First Visit
- Major curve Cobb angle
expanded:
- Age
- Gender
- Prior Spine Surgery
- ASA classification
- Decompression
- Osteotomy
- 3CO
- SPOs
- BMI_First Visit
- Tobacco use_First Visit
- Osteoporosis / osteopenia
- Levels Previously operated - Lower
- LGap
- RLL
- Number of Interbody Fusions
- 'Posterior Instrumented Fusion: Upper / Lower Levels'
- Alif
- LL-Lordosis Difference
- Weight (kgs)_First Visit
- Height (cm)_First Visit
- Total surgical time st1+st2+st3
- Alcohol/drug abuse
- Anemia or other blood disorders
- Osteoarthritis
- Mild vascular
- Depression / anxiety
- Diabetes with end organ damage
- Cardiac
- Hypertension
- Chronic pulmonary disease
- Nervous system disorders
- Renal
- Peripheral vascular disease
- Psychiatric / Behavioral
- Peptic ulcer
- Bladder incontinence
- Bowel incontinence
- Leg weakness
- Loss of balance
- NRS back - Leg pain - Average
- Years with spine problems
- ODI - Score (%)_First Visit
- SRS22 - SRS Total score_First Visit
- SF36 - PCS_First Visit
- SF36 - MCS_First Visit
- Major curve Cobb angle
- SRS22 - SRS Subtotal score_First Visit
- T1 Sagittal Tilt
- Sagittal Balance
- Global Tilt
- Lordosis (top of L1-S1)
- Pelvic Tilt

Age

Proportion of na: 0%

Gender

Proportion of na: 0%
Female Male
No 342 91
Yes 47 16

Prior Spine Surgery

Proportion of na: 0.2%
NA No Yes
No 1 274 158
Yes 0 31 32

ASA classification

Proportion of na: 0.6%

Decompression

Proportion of na: 0%
No Yes
No 240 193
Yes 38 25

Osteotomy

Proportion of na: 0%
No Yes
No 209 224
Yes 20 43

3CO

Proportion of na: 0%
No Yes
No 382 51
Yes 47 16

SPOs

Proportion of na: 0%

BMI_First Visit

Proportion of na: 1.8%

Tobacco use_First Visit

Proportion of na: 2.6%
Current Ex-User NA Non-User
No 74 84 11 264
Yes 11 14 2 36

Osteoporosis / osteopenia

Proportion of na: 0%
No Yes
No 340 93
Yes 54 9

Levels Previously operated - Lower

Proportion of na: 63.1%
C Iliac L NA S T
No 6 8 91 279 43 6
Yes 0 4 16 34 8 1

LGap

Proportion of na: 2.6%

RLL

Proportion of na: 2.6%

Number of Interbody Fusions

Proportion of na: 0%

Posterior Instrumented Fusion: Upper / Lower Levels

Proportion of na: 0%
Iliac+S L T
No 301 128 4
Yes 52 11 0

LL-Lordosis Difference

Proportion of na: 2.6%

Propensity Scores Common Support

## Loading required package: lattice
## 
## Attaching package: 'lattice'
## The following object is masked from 'package:boot':
## 
##     melanoma

Model Stats

  • Treatment proportion: 0.127
  • Model Type: elastic_net
  • Accuracy: 0.8941321
  • Params: alpha: 0.1 lambda: 0.0023528

Model Coefficients

Bootstraping replicas: 50

Average Treatment Effects - Quality Life

Outcome: 2Y. ODI - Score (%)
Distribution:
  0%  25%  50%  75% 100% 
 -67  -27  -14   -4   40 
Model Type Y: boosting 
RMSE: 17.1040302489693 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 18.3268680963959 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): 8.45 (1.827)
Trimmed ATE (Yes-No): 8.628 (1.94)
Upper ATE (Yes-No): 3.837 (2.161)
Observational differences in treatment 0.206 (Yes-No) 

   treatment   outcome
1:       Yes -14.61290
2:        No -14.81887
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 2Y. SRS22 - SRS Subtotal score
Distribution:
    0%    25%    50%    75%   100% 
-0.950  0.215  0.700  1.160  3.050 
Model Type Y: boosting 
RMSE: 0.678801233444384 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 0.689437386123 
Params: nrounds: 50.0
max_depth: 1
eta: 0.4
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): 0.167 (0.132)
Trimmed ATE (Yes-No): 0.171 (0.137)
Upper ATE (Yes-No): 0.052 (0.101)
Observational differences in treatment 0.18 (Yes-No) 

   treatment   outcome
1:       Yes 0.8662500
2:        No 0.6858672
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 2Y. SF36 - MCS
Distribution:
    0%    25%    50%    75%   100% 
-33.82  -3.69   3.72  12.94  39.74 
Model Type Y: boosting 
RMSE: 18.897114325859 
Params: nrounds: 50.0
max_depth: 1
eta: 0.4
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 12.429220721707 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -3.64 (1.852)
Trimmed ATE (Yes-No): -4.032 (1.923)
Upper ATE (Yes-No): 7.01 (2.533)
Observational differences in treatment -0.242 (Yes-No) 

   treatment  outcome
1:       Yes 3.941111
2:        No 4.183228
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 2Y. SF36 - PCS
Distribution:
    0%    25%    50%    75%   100% 
-18.94   0.72   6.64  12.42  38.99 
Model Type Y: boosting 
RMSE: 9.4997068334648 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 9.48733479462741 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): 1.207 (0.959)
Trimmed ATE (Yes-No): 1.642 (1.002)
Upper ATE (Yes-No): -10.579 (1.907)
Observational differences in treatment 1.02 (Yes-No) 

   treatment  outcome
1:       Yes 7.687037
2:        No 6.666929
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Average Treatment Effects - Radiology

Outcome: 6W. Major curve Cobb angle
Distribution:
     0%     25%     50%     75%    100% 
-72.000 -20.510 -10.000  -3.905  30.800 
Model Type Y: boosting 
RMSE: 20.0129644268323 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 13.406769350796 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -0.128 (0.835)
Trimmed ATE (Yes-No): 0.066 (0.875)
Upper ATE (Yes-No): -4.892 (2.765)
Observational differences in treatment -1.878 (Yes-No) 

   treatment   outcome
1:       Yes -14.77412
2:        No -12.89612
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 1Y. Major curve Cobb angle
Distribution:
    0%    25%    50%    75%   100% 
-64.00 -22.69 -10.36  -3.00  22.44 
Model Type Y: boosting 
RMSE: 18.2961436679062 
Params: nrounds: 50.0
max_depth: 1
eta: 0.4
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 14.061345826842 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): 0.318 (2.026)
Trimmed ATE (Yes-No): 0.626 (2.059)
Upper ATE (Yes-No): -6.347 (2.303)
Observational differences in treatment -2.136 (Yes-No) 

   treatment   outcome
1:       Yes -15.53875
2:        No -13.40321
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 6W. T1 Sagittal Tilt
Distribution:
        0%        25%        50%        75%       100% 
-23.631420  -6.000000  -1.411482   1.689195  18.000000 
Model Type Y: boosting 
RMSE: 7.04595548091397 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 5.86250663286041 
Params: nrounds: 50.0
max_depth: 1
eta: 0.4
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -3.917 (0.351)
Trimmed ATE (Yes-No): -3.927 (0.376)
Upper ATE (Yes-No): -3.642 (0.789)
Observational differences in treatment -2.954 (Yes-No) 

   treatment   outcome
1:       Yes -4.965070
2:        No -2.010682
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 1Y. T1 Sagittal Tilt
Distribution:
        0%        25%        50%        75%       100% 
-30.098675  -5.808565  -2.187195   1.000000  20.000000 
Model Type Y: boosting 
RMSE: 5.73130445298575 
Params: nrounds: 50.0
max_depth: 1
eta: 0.4
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 5.74987167040775 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -3.842 (0.56)
Trimmed ATE (Yes-No): -3.711 (0.579)
Upper ATE (Yes-No): -6.224 (0.757)
Observational differences in treatment -2.397 (Yes-No) 

   treatment   outcome
1:       Yes -4.840364
2:        No -2.442963
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 6W. Sagittal Balance
Distribution:
     0%     25%     50%     75%    100% 
-194.79  -69.00  -26.50    3.96  114.15 
Model Type Y: boosting 
RMSE: 62.8381645593175 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 52.4458976074168 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -35.326 (3.599)
Trimmed ATE (Yes-No): -35.092 (3.914)
Upper ATE (Yes-No): -39.887 (10.927)
Observational differences in treatment -33.252 (Yes-No) 

   treatment   outcome
1:       Yes -63.58620
2:        No -30.33438
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 1Y. Sagittal Balance
Distribution:
     0%     25%     50%     75%    100% 
-237.47  -67.07  -30.52    5.84  109.54 
Model Type Y: boosting 
RMSE: 53.6292188663885 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 51.6472545013203 
Params: nrounds: 50.0
max_depth: 1
eta: 0.4
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -40.753 (3.657)
Trimmed ATE (Yes-No): -38.835 (3.748)
Upper ATE (Yes-No): -75.023 (11.96)
Observational differences in treatment -30.124 (Yes-No) 

   treatment   outcome
1:       Yes -59.74368
2:        No -29.61955
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 6W. Global Tilt
Distribution:
    0%    25%    50%    75%   100% 
-68.62 -17.58  -6.00   1.52 149.41 
Model Type Y: boosting 
RMSE: 15.3512673179218 
Params: nrounds: 50.0
max_depth: 1
eta: 0.4
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 14.1301834344565 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -8.99 (1.488)
Trimmed ATE (Yes-No): -8.969 (1.54)
Upper ATE (Yes-No): -9.502 (2.338)
Observational differences in treatment -10.536 (Yes-No) 

   treatment    outcome
1:       Yes -17.585294
2:        No  -7.049362
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 1Y. Global Tilt
Distribution:
     0%     25%     50%     75%    100% 
-62.630 -16.000  -6.465   1.000  26.000 
Model Type Y: boosting 
RMSE: 14.4954468564537 
Params: nrounds: 50.0
max_depth: 1
eta: 0.4
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 11.5791388932799 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -14.424 (1.641)
Trimmed ATE (Yes-No): -14.391 (1.72)
Upper ATE (Yes-No): -15.098 (3.038)
Observational differences in treatment -10.582 (Yes-No) 

   treatment    outcome
1:       Yes -16.991026
2:        No  -6.409266
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 6W. Lordosis (top of L1-S1)
Distribution:
     0%     25%     50%     75%    100% 
-94.930 -24.045  -9.355   0.140  29.000 
Model Type Y: boosting 
RMSE: 21.0603294820447 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 15.5966418482955 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -6.392 (1.271)
Trimmed ATE (Yes-No): -6.24 (1.289)
Upper ATE (Yes-No): -9.991 (2.239)
Observational differences in treatment -10.943 (Yes-No) 

   treatment   outcome
1:       Yes -21.82942
2:        No -10.88661
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 1Y. Lordosis (top of L1-S1)
Distribution:
    0%    25%    50%    75%   100% 
-94.63 -25.71  -9.00   0.00  23.38 
Model Type Y: boosting 
RMSE: 21.6887486613077 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 15.4608886253133 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -12.305 (2.167)
Trimmed ATE (Yes-No): -11.915 (2.273)
Upper ATE (Yes-No): -20.471 (2.901)
Observational differences in treatment -13.732 (Yes-No) 

   treatment   outcome
1:       Yes -24.84000
2:        No -11.10803
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 6W. LGap
Distribution:
       0%       25%       50%       75%      100% 
-96.12340 -24.28110  -9.06300   0.31715  78.92000 
Model Type Y: boosting 
RMSE: 21.7913597510975 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 17.0334976582635 
Params: nrounds: 50.0
max_depth: 1
eta: 0.4
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -6.235 (2.563)
Trimmed ATE (Yes-No): -6.099 (2.641)
Upper ATE (Yes-No): -9.447 (2.712)
Observational differences in treatment -11.262 (Yes-No) 

   treatment   outcome
1:       Yes -21.70294
2:        No -10.44091
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 1Y. LGap
Distribution:
      0%      25%      50%      75%     100% 
-94.8082 -25.2564  -9.0618   0.1456  22.0800 
Model Type Y: boosting 
RMSE: 22.1510025153698 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 15.5787141794885 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -11.896 (2.955)
Trimmed ATE (Yes-No): -11.609 (3.097)
Upper ATE (Yes-No): -17.829 (4.154)
Observational differences in treatment -13.59 (Yes-No) 

   treatment   outcome
1:       Yes -24.56838
2:        No -10.97887
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 6W. Pelvic Tilt
Distribution:
    0%    25%    50%    75%   100% 
-36.41  -8.33  -2.42   2.00  14.42 
Model Type Y: boosting 
RMSE: 10.5808489686285 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 7.47180589398304 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -3.193 (1.109)
Trimmed ATE (Yes-No): -3.07 (1.154)
Upper ATE (Yes-No): -6.259 (1.181)
Observational differences in treatment -6.345 (Yes-No) 

   treatment   outcome
1:       Yes -9.369216
2:        No -3.024645
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Outcome: 1Y. Pelvic Tilt
Distribution:
    0%    25%    50%    75%   100% 
-26.62  -7.10  -2.14   2.00  23.00 
Model Type Y: boosting 
RMSE: 9.7959586330109 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
RMSE: 6.82315217934552 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.6
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): -7.731 (0.677)
Trimmed ATE (Yes-No): -7.91 (0.697)
Upper ATE (Yes-No): -4.043 (0.784)
Observational differences in treatment -5.775 (Yes-No) 

   treatment   outcome
1:       Yes -8.026750
2:        No -2.251544
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Average Treatment Effects - Complications

Outcome: complication
Distribution:
Proportion 
 0.2966102 
Model Type Y: boosting 
Accuracy: 0.589393939393939 
Params: nrounds: 50.0
max_depth: 1
eta: 0.4
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 
Accuracy: 0.701921973608721 
Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): 0.068 (0.054)
Trimmed ATE (Yes-No): 0.061 (0.057)
Upper ATE (Yes-No): 0.254 (0.058)
Observational differences in treatment 0.069 (Yes-No) 

   treatment   outcome
1:       Yes 0.3571429
2:        No 0.2884615
`geom_smooth()` using method = 'loess' and formula 'y ~ x'

Average Treatment Effects - Nombre Revisions

Outcome: reinterventions
Distribution:
  0%  25%  50%  75% 100% 
   0    0    0    1    6 
Model Type Y: boosting 

Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

Model Type No: boosting 

Params: nrounds: 50.0
max_depth: 1
eta: 0.3
gamma: 0.0
colsample_bytree: 0.8
min_child_weight: 1.0
subsample: 0.5

ATE (Yes-No): 0.169 (0.159)
Trimmed ATE (Yes-No): 0.229 (0.166)
Upper ATE (Yes-No): -1.44 (0.15)
Observational differences in treatment 0.155 (Yes-No) 

   treatment   outcome
1:        No 0.4519231
2:       Yes 0.6071429
`geom_smooth()` using method = 'loess' and formula 'y ~ x'